Rapid detection of acute kidney injury by plasma cystatin C in the intensive care unit

Methods. pCysC and pCr were measured on admission to one of two intensive care units (ICU) and then daily over 7 days. Patients were classified according to the analyte that first increased by either >= 25 or >= 50% above the admission value. The proportion of patients in each class was compared using McNemar's chi-square test. Sustained acute kidney injury (AKI, a >= 50% increase in pCr from baseline for >= 24 h), dialysis and death within 30 days were recorded. The ability of pCysC and pCr on admission to predict sustained AKI, dialysis or death was assessed from the area under the receiver operator characteristic curve (AUC). Results. Of 442 patients, 83 had a >= 50% increase in one analyte, 17 in both and 342 in neither. Comparable numbers for a >= 25% increase were 163 in one analyte, 45 in both and 234 in neither. pCysC increased prior to pCr more frequently than vice versa in both the cohort with a >= 50% increase (P < 0.0001) and with a >= 25% increase (P < 0.0001). pCysC predicted sustained AKI with an AUC of 0.80 [95% confidence interval (CI) = 0.71-0.88]. pCysC and pCr were similarly moderately predictive of death or dialysis with AUCs of 0.61 [95% CI = 0.53-0.68] and 0.60 [95% CI = 0.51-0.67], respectively. Conclusion. pCysC was an effective and earlier surrogate marker of decreased renal function than pCr in a general ICU population.